like 3dsmax lattice modifier, but good?

[super gnu]

got to do some "eiffel tower" style lattice metalwork on a (very) large scale... 3dsmax lattice modifier just doesnt cut it.. junctions and beam rotation issues are a show stopper.

apart from fiddly work with the slice tool, inset faces, delete and shell whats left, any clever plugins or scripts about, which could save me time and headaches?

i thought maybe railclone might be able to do stuff like this, but a) ive not seen anything specific mentioned online, b) ive not got railclone c) its not cheap and d) it has a chunky learning curve which id be starting from zero, which might take longer than just using the above technique.

love to hear suggestions.

[^Lele^]

Houdini, if you know it a little bit, could definitely help (and then ofc you could export.).

[super gnu]

ha.. well if railclone has a learning curve...

ive fiddled with some sim stuff in houdini, but never looked at the modelling side of it..

[^Lele^]

Eheh, yes of course it'd be a tad too steep to pick up now.
It does however have exceptional capabilities when it comes to truly procedural modelling.
See this page (and if you care to, check the Entagma tutorial linked in it) for size.
Maybe you can suffer the feeling of inadequacy (that is a step which with Houdini i have since i first tried it in 1996. Still waiting to move one up. XD) for long enough to get a specific result out?

[super gnu]

thanks Lele, yeah i know very well that chilling realisation that you know nothing about 3d, when dipping toes in Houdini and visiting the various fora. It is however accompanied by a large dollop of satisfaction when something you are making *does what its supposed to* even if its built on the shoulders of a dozen tips and suggestions from people with a phd in making you feel inadequate.

[fixeighted]

I had to buy Railclone for the stadium thing I just finished. I could not have done the job without it and it paid for itself easily.

That said, yes it has a awkward learning curve if you jump in to something quite seemingly tricky immediately. However, what
saved me was that the support is absolutely superb. The guys at Itoo basically talked me through all I needed to do, took my messy partial setups and cleaned them up,
explained what they did to make them work and they did all this almost in real-time over a few days.

There is a caveat though; Railclone is fantastic and can do many many things, however there are still some procedures that it just cannot handle, so depending on specifics it still may not
magically create what you need.

Other than that, I'd love to see the actual geometry you need to create...maybe I'll have an approach which would work without the plugins..maybe not

[super gnu]

something as close to this as possible.

[super gnu]

given the tiny scale of the webbing on most elements, and the fact its likely a background element viewed about 3000 pixels wide, probably a combination of mesh slicing/inset, shell to get main structural elements, then lattice to get small trusswork is probably the best bet. you wont see the crappy joins or wierd 45 degree rotation of the beams at that scale.

[fixeighted]

Aah, I see. To me that looks pretty simple to be honest. Many repeating parts that wouldn't take much time to create.
No need for Railclone at all...could probably put it together in a day or a bit longer, at first glance....maybe less if there some reliable dimensions to work from.

I'm not sure what you mean by all the mesh slicing and insets though. I would simply model the various parts of each main segment
and array them.

[super gnu]

I'm not sure what you mean by all the mesh slicing and insets though. I would simply model the various parts of each main segment
and array them.

i was imagining block modelling the form, then using inset and delete centre faces, to get the basic structural elements... shell to give thickness, use of connect to divide edges up, then lattice to truss-ify it..

you are probably right, im overthinking what is actually quite a repetitive structure, and it would be better to just model the elements and duplicate them.

[fixeighted]

Oh yeah, that method of blocking/deriving is more headache than solution imo, unless it really doesn't matter about the details

Simpler to approach it as it is in reality, as at least then you know it works as it should

I was tempted to just say I'll just build it anyway, as it looks like a fun thing to assemble, but I'm off to buy a new bass today, so my
available fingers will be thouroughly occupied

[super gnu]

Oh yeah, that method of blocking/deriving is more headache than solution imo, unless it really doesn't matter about the details

Simpler to approach it as it is in reality, as at least then you know it works as it should

I was tempted to just say I'll just build it anyway, as it looks like a fun thing to assemble, but I'm off to buy a new bass today, so my
available fingers will be thouroughly occupied

ahaha.. noo a new bass sounds much more interesting than modelling trusswork.

besides then id either have to give you the job, and make no money, pay you a pittance to make a dishonest profit, or tell my client and have them contact you.... i dont like any of these options

[super gnu]

p.s. agreed its a cool structure... Lovell Telescope apparently when built it was the biggest movable telescope in the world, still the 3rd or 4th. and was the only one anywhere capable to tracking sputnik when it launched.

[fixeighted]

LOL no I would not have charged for it...it would have just been something to do today. I don't find these things hard at all, so would have been a pleasure

[^Lele^]

Now that i think of it, Modo as well should be able to allow you for massive, procedural replication.
However, my experience of it is scant, and i have no idea as to its pricing structure.

p.s.: If it's very much at the back, have you thought of not modelling it at all, finding a picture of it with the right lighting conditions, and then slotting it in as a card, perhaps with some faked bump for perceptually smaller details?